A neural network approach to real-time dielectric characterization of materials
β Scribed by R. Olmi; G. Pelosi; C. Riminesi; M. Tedesco
- Publisher
- John Wiley and Sons
- Year
- 2002
- Tongue
- English
- Weight
- 77 KB
- Volume
- 35
- Category
- Article
- ISSN
- 0895-2477
No coin nor oath required. For personal study only.
β¦ Synopsis
Abstract
Artificial neural networks (ANNs) are proposed as an inversion approach in microwave measurement methods of the dielectric characteristics of materials. Reflection and transmission methods require a proper electromagnetic (EM) model of the measurement system, and solutions in terms of the material permittivity are usually in implicit form, requiring an inversion procedure that can be costly in terms of computing time. In such applications, the ANN approach is favorable in that the EM computations are performed during the offβline training phase of the network. As an example of a realβworld application, the ANN approach is tested on a waveguide method for the measurement of permittivity in the X band for which a working solution is yet available, showing that the parameters of interest (permittivity and thickness of the material under measurement) can be obtained in realβtime with a negligible loss of accuracy. Β© 2002 Wiley Periodicals, Inc. Microwave Opt Technol Lett 35: 463β465, 2002; Published online in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/mop.10639
π SIMILAR VOLUMES
An analysis of the performances of the neural-network approach for the geometric and dielectric characterization of buried cylinders is carried out. The neural-network process data are obtained from the time-domain formulation of the electromagnetic scattering problem. This analysis is based on the
In this paper a global design method for associative memories using discrete-time cellular neural networks (DTCNNs) is presented. The proposed synthesis technique enables to realize associative memories with several advantageous features. First of all, grey-level as well as bipolar images can be sto
Polymeric materials are finding increasing application in commercial optical communication systems. Taking advantage of techniques from the field of artificial intelligence, the goal of our research is to construct systems that can computationally design polymer formulations, including polymer optic
We explore the use of feed forward artificial neural networks for determining rotational correlation times from slow motional nitroxide electron spin resonance spectra. This approach is rapid and potentially eliminates the need for traditional iterative fitting procedures. Two networks are examined:
The unit vector s was chosen arbitrarily, and Λ0 z f L. The precision was set to 1.0 = 10 y3 . For each L, the function T was evaluated at 2 L 2 observation points. This is L in accordance with the demands of MLFMA. The results are presented in Table 2. In the table, entries marked ''direct'' refer